Finding relevant documents by converting text to numerical vectors and measuring similarity in that space.